Research Article

Yolo-Based Improvements in Remote Sensing Image Applications

Table 4

Comparison experiment on the HRRSD dataset.

ā€‰Fast RCNNGACL
fast RCNN
Faster RCNNGACL
faster RCNN
YOLOv3YOLOv3-DEYOLOv4YOLOv5YOLOv7

Ship75.074.388.588.578.889.8785.099595.8
Bridge75.176.785.585.679.7889.785.5483.486.9
Ground track field90.089.690.690.790.3290.3489.0490.693.4
Storage tank79.880.488.789.289.1289.6789.829698.6
Basketball court36.742.147.949.760.9272.3967.175.785.3
Tennis court75.077.080.780.889.4190.2288.7792.696.5
Airplane83.385.190.890.890.8490.8890.4394.797.7
Baseball diamond83.682.686.987.273.5488.481.3883.588.4
Harbor76.078.489.489.787.6890.1290.2496.897.8
Vehicle46.150.784.086.988.5989.9783.6191.594.3
Crossroad67.168.788.688.287.4187.8388.9684.487.9
T junction39.238.875.175.073.5373.5566.988489.5
Parking lot37.539.563.365.357.6262.9159.966172.1
Mean AP66.568.081.582.180.5885.06582.07186.861591.092